The theory of granular packings as a chapter of

the soil mechanics subject

Calixtro Yanqui
Department of Civil Engineering San Agustin National University of
Arequipa, Arequipa, Peru

ABSTRACT
Soils are discontinuous substances made of individual solid particles and voids, defined by their contacts. Therefore, the
natural description of soils is better accomplished by using granular packings. In this context, to achieve the maximum
mathematical simplicity, a grain is represented by a sphere whose diameter is equal to the centroidal distance between
two grains in contact, and the fundamental element, by the parallelepiped drawn by the centroides of eight neighbouring
spheres, which must be in statically equilibrium, whose solid volume is obviously equal to the volume of the effective
grain, and whose geometric parameters define the structure of the assembly, as well as the total volume. For the
simplest granular packing, the Wadells shape coefficient and the Hazens uniformity coefficient are redefined. Under
these considerations, it is settled down the fundamental equation that relates any global parameter, such as the void
ratio or the volume ratio, with the structural parameters of the element. This connection is used to explain and calculate
the physical and mechanical properties of soils; for instance, the relationship between the densest state and the loosest
state, the relationship between the angle of internal friction and the coefficient of lateral stress "at rest", the relationship
between that angle and the porosity, and the value of the Casagrandes critical void ratio, among others. All relationships
so obtained fit very well with the experimental data reported by worldwide authors.

RESUMEN
Los suelos son sustancias discontinuas conformadas por partculas slidas y poros, definidos por sus contactos. En
consecuencia, la descripcin natural de los suelos se consigue de una mejor manera usando los ensambles granulares.
En este contexto, para conseguir la mxima simplificacin matemtica, un grano se representa por una esfera cuyo
dimetro is igual a la distancia centroidal entre dos granos en contacto y el elemento fundamental, por el paralelelppedo
dibujado por los centroides de ocho esferas vecinas, las cuales deben estar en equlibrio esttico, cuyo volumen es
obviamente igual al volumen del grano efectivo, y cuyos parmetros geomtricos definen la estructura del ensamblaje,
as como el volumen total. Para una ensamble granular simple, se redefinen el coeficiente de forma de Wadell y el
coeficiente de uniformidad de Hazen. Bajo estas consideraciones, se establece la ecuacin fundamental que relaciona
cualquier parmetro global, como el ndice de poros o el indice volumtrico con los parmetros estructurales del
elemento. Esta relacin se usa para explicar y calcular las propiedades fsicas y mecnicas del suelo; por ejemplo, la
realacin entre los estados ms denso y ms suelto, la relacin entre el ngulo de rozamiento interno y el coeficiente de
esfuerzo lateral en reposo , la relacin entre aquel ngulo y la porosidad, y el valor del ndice de poros crtico de
Casagrande, entre otras. Todas estas relaciones as obtenidas se ajustan bien a los datos experimentales reportados
por autores de todo el mundo.

1 INTRODUCTION outlined in this report is applicable to fine soils as well, to

Soil is inherently a particulate system. Indeed, the
science that treats the stress-strain behavior of soil may
well be thought of as particulate mechanics (Lambe and
Whitman, 1969). The way out of the difficulty lies in
dropping the old fundamental principles and starting again
from the elementary fact that the sand consists of
individual grains. (Terzaghi, 1920). Statements of this
kind have been made several times by prominent authors.
Therefore, it is compulsory to introduce some granular
model in the Soil Mechanics to grasp its particulate
nature. Being a branch of the physical science, this model
must exhibit three merits: comprehensiveness, predictive
power and simplicity (e.g. Brancazio, 1975). Within this
frame of reference, a new chapter of the soil mechanics
subject is proposed to rationally explain the changes of
porosity, the extreme states of compactness, the
transmission of simple stresses, the shear strength, and
the critical state of granular soils. But the model to be
explain quantitatively the flocculation of silty grains, the
polarization of clay sheets, the effect of the adsorbed
double layer, the nature of Atterberg limits, among other
topics that shall not be treated here.
2 THEORY OF GRANULAR PACKINGS
Soil is a discontinuous substance made of an assemblage
of grains and pores. Grains are solid bodies arbitrary in
size, shape, orientation and surface texture. Pores are the
space where there are no grains and may contain air and
water. Grains are interconnected through almost punctual
contacts, forming a highly complex and generally random
system, referred to as soil structure. From the practical
point of view, two features are most important in the
description of the soil: the shape and the gradation of the
grains.
A granular packing is an orderly regular array of
spheres of the same size and smooth surface texture.
Lattice is the arrangement of the centers of the spheres,
called homologous points, and obeys the laws of
symmetry of crystals, for which, Bravais demonstrated, as
early as 1848, that can only have fourteen kinds
physically acceptable. The oblique parallelepiped,
constituted by eight homologous points, neighbors with
each other, pertaining to the lattice is called the unit cell
(Klein and Hurlbut, 1996). In this context, a fundamental
assumption is stated: a soil can be modelled as an ideal
granular packing made of spheres representatives of all
real grains. This transformation can be achieved through
a proper definition of their physical characteristics, called
textural parameters.
2.1 Grain equivalent diameter
The equivalent diameter comes from the most basic
transformation of a grain into a sphere, and has been
proposed by many authors, mention apart that
constitutes the fundamental working hypothesis in
assessing the size of the fine soils, for example, when
using the hydrometer. The equivalent diameter, D, is the
diameter of the sphere of equal volume as the grain, Vs,
(Fig. 1), thus
6 Vs
D 3 [1]
The equivalent diameter of coarse sand and gravel can
be determined using the pycnometer method, whereby
the volume of grain is equal to the volume of water
displaced. The equivalent diameter of fine soil, obtained
by the adsorption test, is defined as the diameter of the
sphere of equal specific surface area, Ss. and equal
weight, s, to the grain
6
D [2]
Ss s representative integral diameter of the soil.
Figure 1. Equivalent diameter and diameter of contact for
grains.
2.2. Contact diameter
The measure of the separation between the centroids of
two grains is called contact diameter, Dc, (Fig.1) and can
be determined for gravels by a simple test. For better
statistical accuracy, several coarse grains are placed in a
channel of semicircular section and the distance between
the first and last grain is measured. This value is divided
by the number of grains involved to get the contact
diameter. Obviously, the grains should be of the same
size and, at least in theory, the contact diameter is always
greater than the equivalent diameter.
2.3. Coefficient of particle shape
Originally, Wadell defined the shape factor or sphericity of
a grain as the ratio of grain surface area and surface of
the equivalent sphere. But, due to practical difficulties,
Wadell himself later amended this definition to the
relationship between the volume of the grain and volume
of the circumscribed sphere. In this theory, this coefficient
is defined as the ratio of the volume of the grain and the
volume of the sphere of contact
3
1 Vs D
a [3]
a Vc Dc
2.4. Generalized uniformity coefficient
The gradation of the soil is the second feature to be
modified. It is said that a soil is uniform when consists of
grains of equal size. To take into account the fact that the
vast majority of soils are, on the contrary, graduated,
Hazen defined the uniformity coefficient as the
relationship between the diameter D60 and the diameter
D10. The first represents approximately the average
diameter of the soil and the second, the effective
diameter. Therefore, the generalization of this factor leads
to the following expression:
3
D
u [4]
Def
where D is the structural diameter and Def, the
2.5. Textural grain coefficient
The two texture coefficients have the same effect on the
packing: the content of pores increases with the angularity
and uniformity of grains and decreases with the
roundness and gradation of sizes. Since in general, it is
not possible to discriminate the participation of each of
them, is more practical to use a single textural grain
coefficient:
g u a [5]
2.6. Unit cell volume
The main merit of the unit cell is the universality of its
volume, because, as it is well known, the volume of a
parallelepiped is found by multiplying the area of the base
by the height. This means that, whatever be the kind of
lattice, the volume is met by knowing the dimensions and
directions of the edges of the parallelepiped formed by
eight neighboring homologous points. In a granular
packing, the contact diameter is a constant quantity that
depends on the textural characteristics of soil. The angle
of the basal parallelogram , and the angle that the
generatrix makes with the vertical line, not only define the
lattice structure, but allow the classification of granular
packings. For example, considering the ease of
representation, they are classified as two and three
dimensional, and if it is considered the nature of the
directrix, in prismatic and pyramidal.

Figure 3. Plan and elevation view of the two-dimensional

Figure 2. The lattice unit cell, and its geometrical packing. a) Equilateral parallelogram in the OXZ plane
elements. and squares in the OXY plane. b) Isosceles parallelogram
in the OXZ plane.
2.6.1. Two-dimensional packings

Although in essence they are three dimensional, these

packings can be represented in two dimensions due to its
symmetry with respect to one of the Cartesian vertical
planes. This means that the horizontal face and the
oblique face are squares of side equal to the diameter of
contact, while the vertical face is a parallelogram defined
by the angle of the generatrix. In the nomenclature of
Bravais, this lattice is called monoclinic, and geometrically
corresponds to a parallelogram, which can be classified
according to the location of the diameters of contact as
equilateral and isosceles.

2.6.1.1 Equilateral parallelogram

In this case, all sides of the parallelogram are diameters

of contact, so the volume is expressed in terms of the
angle that the oblique side makes with the vertical, . Figure 4. Plan and elevation view of the three-dimensional
3
Then: V=Dc cos. packing. a) Equilateral parallelepiped, b) Isosceles
tetrahedral parallelepiped.
2.6.1.2. Isosceles parallelogram

In this type of packing, the horizontal sides of the
parallelogram are not diameters of contact, owing to
which a grain from the upper base must rest on the two
grains of the lower base. So that, the oblique side is equal
to the minor diagonal of the parallelogram and the angle
with respect to the vertical is denoted by . Therefore,
3
V=Dc sin2.
2.6.2. Three-dimensional packings
In the most general configuration, three-dimensional
packings require for its description of at least two angles.
In the nomenclature of Bravais, this lattice is called
triclinic and geometrically corresponds to a parallelepiped,
which, like two-dimensional packings, can be classified as
equilateral and tetrahedral.
2.6.2.1 Equilateral parallelepiped
In this packing, all edges of the parallelepiped are
diameters of contact, and each grain automatically
satisfies the condition of static equilibrium. In this case,
the angle of the generatrix with the vertical is denoted by
and the total volume is given by the expression:
3
V=Dc sincos. According to the angle , the base of this
parallelepiped varies from a hexagonal rhomb to a
square.
2.6.2.2 Tetrahedral parallelepiped
The directrix of this packing does not consist of contact
diameters, and thus, every grain of the top layer is not in
equilibrium, except if is supported by two grains of the
layer below. When this occurs, the diagonals of the lateral
faces of the parallelepiped are equal to the oblique edge 90- and so on. Likewise, the equivalence of the two-
and the parallelepiped is symmetric with respect to the dimensional packings themselves points out the following
plane containing the angle of the generatrix with the relationships:
3
vertical. Then: V=Dc sin(1+cos)sinsin2. According to
the angle , the base of this isosceles tetrahedron varies 45 /2 or 45 /2 [12]
from an equilateral triangle to a square. In the last case,
the tetrahedron transforms itself to an octahedron. These facts illustrate one of the intrinsic properties of
granular packings: their duality or, in a more general
2.7 Solid phase volume in the unit cell sense, their multiplicity.

The second merit of the unit cell is the constancy of the 3.1 Extreme states
volume of solids. It is also a known fact of elementary
geometry that the sum of the eight octants formed by The most important conclusion of the analysis developed
three oblique planes is equal to the total solid space, up to this point is that the porosity of the granular
regardless of the angles and . This principle also packings changes according to the value of or and .
applies to a sphere and, even more, for any solid body. The second conclusion concerns the restrictions imposed
Indeed, the sum of the eight spherical trihedrons, defined by the contact between grains, by which the porosity is a
by the faces of the parallelepiped is equal to the integral bounded quantity, called minimum porosity, n m, and
effective volume of the grain: related to the densest state of the packing. In the two
kinds of three-dimensional packing, the angle can only
take values between 60 and 90. Furthermore, in the
Vs Def 3 [6]
6 pyramidal packing, these extreme values are related to
two axisymmetric lattices: the tetrahedral, for = 60 ,
3 PHASE RELATIONSHIPS and the octahedral for = 90 . Likewise, the angles
and , called structural angles, are bounded. In summary,
Once the total volume and solid volume are known, the the following ranges of validity are recognized: a) for the
amount of pores that contain the packing may be equilateral parallelogram: -30 30; b) for the
determined. Historically, different definitions have been parallelogram isosceles: 30 60; c) for the
proposed, according to the need of the subject, for equilateral parallelepiped: -1/2secsin1/2sec;
example, the porosity, which relates the pore volume to 0/2; 6090; and d) for a tetrahedral
the total volume, n = Vv/V, the void ratio, e, which relates parallelepiped: arcsin[1/2sec(/2)], 6090.
the pore volume with the volume of grains, e = V v/Vs, or,
more recently, the volumetric ratio, v, which relates the Table 1. Features of granular packings.
total volume to the volume of grains v = V/Vs, which leads Packing Tetrahedral Octahedral
to a greater compactness of the formulas (Wood, 1990). Loosest 1 1
State cos cos
As they are all different expressions of the same thing, 3 3
these parameters are related to each other: Densest 1 1
State sin sin
1 3 2
e 1 [7]
1 n Minimum 2 2
porosity nm 1 nm 1
6 g 6 g
Hence, the volumetric ratio assumes a definite form for Maximum
nM 1 3
each type of packing: porosity 6
nM 1
g 8 g
6 Extreme nm 1.41nM nm 1.089nM
Equilateral parallelogram: g cos [8]
porosities 0.4144 0.087
Relationship
6
Isosceles parallelogram: g sin 2 [9] The third conclusion is related to the fact that each
equation of the porosity accepts a mathematical root,
which are related to the loosest state of the soil, and,
6
Equilateral parallelepiped: g sin cos [10] therefore, to the maximum porosity, nM, which separates
the acute configuration from the obtuse configuration,
Tetrahedral parallelepiped: and, therefore, is unique for each packing. The respective
6 derivation yields the following results: = 0, for the
g sin (1 cos )
sin sin 2 [11]
equilateral parallelepiped; = 45, for the parallelogram
isosceles and = arccos(1/3) 54.74 for the
A quick inspection of these equations leads to the tetrahedral parallelepiped. The substitution of these
conclusion that, due to the nature of trigonometric values and the corresponding extreme values for the
functions, all of them accept two values. For example, the densest state in equations (8) and (11) allows to
first equation is valid for and -, the second for and establishing a relationship between the maximum porosity
and the minimum porosity for each type of packing. In
table 1, all values for the limit packings of the tetrahedral
family are shown. Just for the sake of identification, they
are named tetrahedral and octahedral
4. VALIDATION OF THE GRANULAR MODEL
The merit of the extreme states lies in the fact that they
can be found experimentally by very simple testing. The
maximum porosity is determined by the uniform pouring of
the grains into a calibrated mold, and the minimum
porosity, by the strongly penetration of a thin rod into the
soil in the mold, by compacting it with a vertical hammer,
by shaking it onto a vibrating table, or combining some of
these procedures. In Figure 5, it is shown the
experimental values for granular soils reported by various
authors (Selig and Ladd, 1973), which can be compared
with the theoretical values given by the extreme
conditions for the tetrahedral and octahedral packings. It
is worth to observe that the experimental data are in the
domain bounded by the lines for the two axisymmetric
pyramidal packings. The second conclusion drawn from
this diagram is that the extreme porosities of a real
granular substance are a measure not only of the pore
volume but also of the grains shape and gradation: g. To
the same conclusion arrived Talbot and Richart, as early
as 1923, based on numerous experimental tests to obtain
the densest state of coarse aggregates.
50
40
Minimum Porosity (%).
1.2
1.1
30
= 1.0 1
20
90 0.9
10
80
70 60 0.8
0 the spheres of contact, the displacement caused by the
20 30 40 50 60
Maximum Porosity (%)
Figure 5. Maximum porosity versus minimum porosity
diagram. The experimental data were reported by several
authors. The lower straight line stands for the tetrahedral
packing, and the higher, for the octahedral one.
5. MECHANICS OF SIMPLE GRANULAR PACKINGS
Unlike continuous media, the granular packings are prone
to exhibit multiple mechanisms of transmission of the
stress, depending on the nature of these and the contact
points between grains. In the last two decades, random
functions of the quantum mechanics have been used with
some success to solve some particular problems (eg,
Aste et al., 2002). Solutions so found are complicated and
inaccessible from the point of view of engineering
practice. However, the application of the principle of the
mean value allows the deterministic and simple
calculation of stress distribution under a general
solicitation and boundary conditions (Yanqui, 1995). But
the analysis of stress in a specimen subject to a uniform
solicitation becomes an extremely simple task, if one
accepts the principle of centroidal reactions (Trollope,
1956; Yanqui, 1980), whereby the contact lines of the unit
cell coincide with the directions of contact forces. Some
authors (e.g. Ostojic, 2006) have called this network of
centroidal reactions a force network ensemble, which,
however, does not necessarily coincide with the unit cell.
Figure 6. Mechanics of the shear stress. a) Dilatant
element, b) Contractive element.
5.1. Simple shear
The prismatic packing is the best model to describe a
simple shear test. In this case, all the edges of the unit
g = cell are diameters of contact and, therefore, the forces
1.3 ensemble coincides with the granular lattice (Fig. 6). The
resultant of the vertical normal force N and the horizontal
force T that supports a grain should be fully transmitted to
the corresponding point of the lower layer, as long as in
an element that works exclusively by shear the normal
component in the horizontal edge is zero. Therefore, if the
resultant coincides with the generatrix of the
parallelepiped, this is a shear element that responds by
diagonal compression. Another important aspect is the
deformation of the granular packing. Being relatively rigid
70 horizontal shear has a horizontal and a vertical
component. The latter is related to so-called "dilatancy" of
granular soils, analyzed first by Reynolds in 1885. But,
because of the dual nature of the packing, also the
opposite phenomenon may occur, which will be called
"contractancy." In conclusion, considering the vertical axis
OZ directed downward, the packing is dilatant when the
sign of is positive and the configuration is acute, while
the packing is contractive when the sign of is negative
and the configuration is obtuse.
5.2 Two-dimensional compression
The two-dimensional simple compression test is
described by a rhombic ensemble of forces by
consideration of the horizontal symmetry (Fig. 7). In this
case, the granular packing does not match the ensemble
of forces, but the assemblage angle is the same for
both, as well as the porosity. The principle of the
centroidal reactions provides that a vertical force P
applied to a grain of the upper layer is divided into two
contact forces, F, symmetrical and oblique, whose
magnitude is given by the expression: F = P / (2cos).
Therefore, the equilibrium in the horizontal direction
requires a horizontal force of magnitude Q = Ptan, or, in
terms of stresses:
3i 1 tan2 [13]
where 1 is the vertical stress and 3i, the internal
horizontal stress, which must be compensated by
applying a confining lateral stress of the same magnitude
at least. This stress makes the difference with the
continuum, which obeys the Cauchys principle, for which
the stress is transmitted as if the substance were
composed by wires parallel to the axis and independent of
each other. In a granular medium, this mode of
transmission of the stress is simply impossible because
the pores of the packing generate oblique contact forces.
Also, due to the dual nature of granular assemblies, there
are two complementary values of the angle for the same
void ratio. For less than 45 , the contact lines coincide
with the force ensemble. For greater than 45 , the
active state is possible only if the contact forces are not
centroidal. Cinematically, in the first case, the element is
dilatant and, in the second, contractive.
Figure 7. Mechanics of the two-dimensional confined
compression. a) Contractive force ensemble, b) Dilatant
force ensemble.
5.3 Triaxial compression
Regarding the axisymmetric character of this test, two
force ensembles are possible: the rhombohedral and the
octahedral, both equilaterals. In this case, the force
ensembles do not coincide with the unit cell, except the
angle, and the porosity. In this ensemble, the axial force
P, acting at a grain, is decomposed into N contact forces,
F = P / (Ncos), where N = 3 if it is a rhombohedron, and
N = 4, if it is an octahedron (Fig. 8) In the first case, the
problem is isostatic and in the second, hyperstatic. The
transformation of this contact force to an average stress
allows the calculation of the internal lateral stress:
tan 2
3i 1 [14]
2 6.2. Two-dimensional compression
that is balanced by the lateral pressure applied to the
specimen. Although the analysis is more complicated in
this case, the dual character of the granular packing also
leads to define it as dilatant if is less than 54.47, and
contractive if is greater than 54.47.
Figure 8. Mechanics of the three-dimensional stress. a)
Simple shear ensemble b) Triaxial compression
ensemble.
6 STRENGTH OF GRANULAR PACKINGS
6.1. Simple shear
At the time of the failure of a granular packing of acute
configuration by horizontal shear, the angle of the
resultant with the normal force equals the angle of internal
friction, , and therefore: = . In a granular packing of
obtuse configuration, the coincidence of the centroidal
reaction with the generatrix is impossible. But, since the
force resultant of the top layer should be transmitted in
some way to the layer below, the only rational possibility
is the appearance of a shear force at the contact, which,
at failure, is equal to the average shear strength between
the grains, and, hence, independent of its assemblage.
Then, = cv. In the first case, the angle of internal
friction depends solely on the structure of the packing.
Mechanically, this is only possible if the deformation is
very small, and grains rotate from one another, as has
been proposed by some authors (e.g. Skinner, 1969). But
this mechanism is also physically impossible, unless the
affected area be a narrow band that acts as a hinge
between the two unaffected portions of the specimen.
This is an experimental fact recognized since the
beginning of soil mechanics (e.g. Taylor, 1948). Some
authors argue that this band has a thickness of about ten
times the diameter of the grain, based on the X-Rays
analysis (e.g. Budhu, 2000). In the second case, the state
of failure is reached when the grains have slipped enough
respect to the neighboring grains. Consequently, the
entire mass is involved, and the required deformation is
relatively large. These findings are also well known from
the experimental results.
The two failure mechanisms described above have a
common point, in which the packing does not expand or
contract itself. Indeed, Casagrande (1936) found
experimentally that there was a value of the void ratio for
which granular soils were strained at constant volume,
and called it critical void ratio.
 the relationship between K0 and is achieved, which also
At failure, the following conditions hold: = , 1 = 1f is not unique. However, further analysis shows the values
and 3i = 3f . Then, according to equations (12) and (13), of K0 bounded as follows:

1 1
f 45 , 3f 1f tan2 f [15] 2
K0 2
[17]
2 1 3 1 2

which coincides with the Mohr-Coulomb law for granular where =tan is the coefficient of internal friction. Figure
soils. In this equation, f represents the plane of failure 9 shows that the extensive data gathered by various
and coincides with the line of contact, as it should be. authors are well suited to this band, and that the lower
Having recognized this equivalence, all of the features limit approaches the empirical formula proposed by Jaky
deduced for the parallelogram packing are valid for the (1944).
rhombic packing.
1
6.3. Triaxial compression
0.9 Moore
In this case, it is not certain that there is a simple 0.8 Kezdi
equivalence between the prismatic force ensemble, c) Hendron
0.7
describing the shear deformation in soils, and the b) Edil
rhombohedral or octahedral force ensembles, which 0.6 Krizek
represents the triaxial test. This is due to the effect of the 0.5 Al-Hussaini
second angle of assemblage . However, a simple

Ko
0.4
qualitative analysis allows for deducing that any
conclusion drawn for the two-dimensional force ensemble 0.3
is valid for three-dimensional force ensemble; for 0.2
instance, the specimen fails along an oblique face of the a)
0.1
polyhedron, when the soil is dense, regarding a dilatant
0
deformation process, and it fails in a bulk fashion when
0 10 20 30 40 50 60 70 80 90
the soil is loose and undergoes a contractive straining.
Friction angle ()
But the derivation of the failure law for this test shall be
done using another route.
Figure 9. Relationship between the internal friction angle
and the coefficient of lateral stress at rest. a) Lower
7 RELATION BETWEEN AND KO limit, b) Upper limit, c) Jakys empirical relationship.

The most basic but very important application of the

principle of the centroidal reactions in dense soils, whose
contact forces coincide with the direction of the contact
lines, is the determination of the natural stress state of the
subsoil. For instance, if a semi-infinite soil that extends
indefinitely in depth and is limited at the top by a gently
sloping surface is considered, the stress components are
obtained by adding all the contact forces acting along
each line of contact passing through the grain considered,
which depends on the type of packing. For a three-
dimensional problem, there are three or four directions,
according to the packing, that can be rhombohedral or
octahedral. Immediately, it shows up that the stress state
is given by equation (14) for a horizontal surface; and,
therefore, the coefficient of lateral thrust "at rest", K 0, is
expressed as:
tan2
K0 [16]
2 figure 10 to show their good correlation with the
But when the surface is inclined, stresses vary according
to the orientation of the base of the ensemble. However,
whatever the orientation be, for a slope angle equal to the
angle of internal friction, , stresses into the subsoil must
meet the criterion of Mohr-Coulomb failure. Removed the
angle from these two conditions of the surface plane,
8 RELATIONSHIP BETWEEN AND POROSITY
The connection between the granular packing and the
force ensemble in the triaxial compression test is the
structural angle . If Ko is removed from equations (16)
and (17), the relationship between the structural angle
and the internal friction angle is obtained. Likewise, If is
eliminated from equation (11), the following relationship is
attained for the lower limit of K0 :
6 g
cos 2 3 2 cos 2 [18]
This equation has been compared with the results of
careful tests carried out and reported by several authors.
Just for illustration, two well known data are presented in
theoretical curves withdrawn from equation (18) for the
dense state, and the horizontal line, = cv, representing
the critical friction angle, for the loose state. The
intersection of this two lines gives the Casagrandes
critical void ratio.

40
Friction angle ()
35
30
0.5 0.6 0.7 0.8
40
Friction angle ()
35
30
0.5 0.6 0.7
Initial void ratio
Figure 10. Relationship between the initial void ratio and
the friction angle. a) for Brasted sand, g=1.016,
(Cornforth, 1973) , b). for medium fine sand, g= 0.993,
(Rowe,1962),
CONCLUSIONS
To be useful for the theory of granular packings, the
Wadells shape coefficient and the Hazens uniformity
coefficient should be redefined. The appropriate choice of
the lattice unit cell generalizes the total volume and the
effective solid volume of the packing. All kinds of packings
may be reduced to only three. There is a definitive
relationship between the minimum and the maximum
porosity in granular soils. Packings exhibit a dual
character at least. Dense packings obey the principle of
the centroidal reactions in studying the stress
transmission under simple boundary conditions, such as
soil testing. Also, the granular packing reveals in a simple
manner the critical state behavior of the shear strength of
soils. In general, the granular packing allows for grasping
and deepening into the particulate nature of soils, and the
results fit well with the experimental data. Therefore, a
granular packing, so defined, is a simple, deterministic,
realistic, and quantitative model for soils, so that it is the
time to include it as a chapter of Soil Mechanics subject.
REFERENCES
Brancazio, P.J. 1975. The Nature of Physics. Macmillan
Publihising Co., Inc. New York.
Budhu, M. 2000. Soil Mechanics and Foundations. John
Wiley & Sons, New York.
Casagrande, A. 1936. Characteristics of cohesionless
soils affecting the stability of slopes and earth fills. J.
Boston Soc. Civil Engs. Vol. 23: 13-32.
Cornforth, D.H. 1973. Prediction of drained strength of
sands from relative density measurements. ASTM
STP 523 : 281-303.
Edil, T.B. and Dhowian, A.W. 1981. At-rest lateral
pressure of peat soils. J. Geotech. Eng. Division,
ASCE, Vol. 107: 201-217.
0.9 1 Hendron, A.J. 1963. The behavior of sand in one-
dimensional compression. Ph.D. Thesis, University of
Illinois.
Kezdi, A. 1966. Contribution to the investigations of
granular systems. Rheology and Soil Mechanics.
Springer-Verlag. Berlin.
Klein, C. and Hurlbut, C.S. 1996. Manual of Mineralogy,
21th Edition. John Wiley & Sons, New York.
Lambe, T.W. and Whitman, R.V. 1969. Soil Mechanics.
John Wiley & Sons, New York.
Moore, C.A. 1971. Effect of mica on Ko compressibility of
two soils. J. Soil Mech. Found. Division, ASCE, Vol.
97: 1275-1291.
Ostojic, S. 2006. Statistical Mechanics of Static Granular
Matter. Ph. D. Thesis, university of Amsterdam.
Rowe, P.W. 1962. the stress-dilatancy relation for static
0.8 0.9
equilibrium of an assembly of particles in contact.
Proc. Royal Soc. , A269: 500-527.
Selig, E.T. and Ladd, R.S. (Eds.). 1973. Evaluation of
Relative Density and its Role in Geothecnical Projects
Involving Cohesionless Soils. ASTM, STP 523.
Skinner, A.E. 1969. A note on the influence of interparticle
friction on the shearing strength of a random
assembly of spherical particles. Geotechnique, No. 19:
150-157.
Taylor, D.W. 1948. Fundamentals of Soil Mechanics.
John Wiley & Sons: New York.
Terzaghi, K. 1920. Old earth pressure theories and new
test results. Engineering News Record, vol. 85.
Trollope, D.H. 1956. The Stability of Wedges of Granular
Materials. Ph. D. Thesis. University of Melbourne.
Wood, D.M. 1990. Soil Behaviour and Critical State Soil
Mechanics. Cambridge University Press. Cambridge.
Yaky, J. 1944. The coefficient of earth pressure at rest. J.
Soc. Hungarian Architecs Eng. Vol. 7, 355-358.
Yanqui, C. 1980. Clastic stresses in infinite slopes. Proc.
rd
3 Peruvian Conference on Civil Engineering: 234-
248. Cusco. Peru.
Yanqui, C. 1982. Statics of gravitating discontinua. MSc.
thesis, The University of South Carolina. USA.
Yanqui, C., 1995. Stresses induced by footings in
th
discontinuous media. 10 Panamerican Conf. on Soil
Mechanics and Foundations. Vol. 3: 1341-1353.
Guadalajara, Mexico.